深層学習を利用した前腕の縦断超音波画像群からの表皮・血管推定と3次元モデル生成

抄録

In this study, we developed an automatic detection algorithm of vessel and skin regions in a transversal ultrasonography image on the arm. We also developed an algorithm to generate 3D model from detected areas for training and assisting nurses' puncture on vein. In the algorithm, the candidate regions of vessel were detected using U-net, which is a kind of deep learning method for segmentation, and then appropriate regions were selected based on vessel properties. The skin regions were also detected using U-net. The 3D polygon data was created from paired pixels in sequential images. The experiments based on single arm scan data demonstrated that our developed model have capablity to detect vessel and skin regions and feasibility to confirm blood vessel running under arm surface.